Fractures and resulting morbidity, mortality and costs are a large and growing burden among older men. The fracture rate increases exponentially after age 65, and the average age of hip fracture is >80 yrs. Fracture risk in later life is strongly influenced by long term declines in bone density, structure and strength, and during this period skeletal change accelerates. Yet, in older men there are very few long-term longitudinal data concerning these changes and how they contribute to the very high risk of fracture during this critical time. Whereas low hip and spine BMD by DXA are strong risk factors for fracture, most fractures occur in those without BMD-defined osteoporosis. The association between change in DXA BMD and incident fracture is controversial, in part due to insufficient long-term data. Moreover, changes in bone strength, assessed by CT- based finite element analysis (CT-FEA), are only modestly associated with changes in DXA BMD. The nature of long-term changes in bone mass and structure, and how they relate to bone strength and fracture risk, remain ill-defined. Furthermore, whereas hip and spine fractures have received the most attention, non-hip, non-vertebral fractures are an enormous medical and societal burden. Our preliminary data show that HR- pQCT measures of bone microstructure and strength at the peripheral skeleton predict fracture independently of DXA BMD. However, the nature of age-related declines in bone strength at the peripheral skeleton is poorly understood, particularly in the oldest men in whom fractures are most likely. By virtue of its comprehensive bone imaging, long duration and very old age of the current cohort, MrOS provides a unique opportunity to address these issues. We propose additional measures in MrOS, and further follow-up for additional fracture ascertainment, enabling a systematic examination of how long-term change in critical skeletal measures influence fracture risk in men during the period of life when the rates of bone loss are most rapid and fracture risk is highest. We will address several aims, including three focusing on hip and spine bone density, structure and strength: 1a) Characterize long-term change in hip and spine bone density, structure and strength over 20 years; 1b) Determine whether baseline BMD and long-term change in BMD independently predict incident non- vertebral fractures; 1c) Determine whether bone strength by CT-FEA predicts incident non-vertebral fractures. In addition, we will address three aims regarding peripheral bone density, structure and strength, including: 2a) Determine the association of HR-pQCT measures with hip and non-hip non-vertebral fracture; 2b) Establish the rate and character of peripheral bone loss in older men; and 2c) Assess the contribution of key determinants of bone strength to loss of peripheral bone density, structure and strength. By leveraging the unique advantages of MrOS, we can advance the understanding of long-term skeletal change, improve fracture prediction and provide new opportunities for prevention of fracture in men.